Oct. 9, 2024
Comments from Center Directors and others regarding the 2024 Nobel Prize in Physics
We would like to extend our heartfelt congratulations to Dr. John Hopfield and Dr. Geoffrey Hinton upon winning the Nobel Prize in Physics. The following are messages from RIKEN Center Directors and others.
- Masashi Sugiyama, Director, RIKEN Center for Advanced Intelligence Project (AIP)
The awarding of the Nobel Prize in Physics to Professors John Hopfield and Geoffrey Hinton is a historic event that gives machine learning recognition as a field of scientific research.
I hope that this will be an opportunity for machine learning research to develop even further and contribute even further to the happiness of humanity.
- Naonori Ueda, Deputy Director, RIKEN Center for Advanced Intelligence Project (AIP)
I would like to offer my heartfelt congratulations to Professors John Hopfield and Jeffrey Hinton for winning the Nobel Prize in Physics. Professor Hopfield proposed a model known as the Hopfield network,1 and with it laid the theoretical foundations for neural networks by combining physics and neuroscience. Professor Hinton, for his part, made breakthroughs in the field of deep learning, and made significant contributions to the technical development of multilayer neural networks, including back-propagation.
Having the opportunity to study in Professor Hinton's laboratory at the University of Toronto and the University of London when I was young was an invaluable experience for me. I still have vivid memories of those days and in particular, I still fondly remember the days when we had fun chatting over lunch, especially at times when he would give us creative quizzes rather than discussing research. His sense of humor and flexible thinking stimulated my curiosity and became a driving force for my studies.
Professor Hinton proposed a scheme for generative models, as typified by Boltzmann machines, and played an important role as a pioneer in this field. I myself was influenced by this and have continued to work on generative models to this day. When he received the NEC C&C Prize and Honda Prize a few years ago, Dr. Hinton was kind enough to ask me to give a guest speech, and he also visited our center to give a special lecture. The fact that he has been so kind to me over the years is a great treasure for me.
I have always admired not only Professor Hinton's outstanding achievements as a researcher, but also his warmth and depth as a human being. The Nobel Prize in Physics that they have both received this time is the result of the wide recognition of their achievements over many years, and is a well-deserved honor. I pray for the continued good health of both, and once again express my deep respect for their great achievements to date.
- Makoto Taiji, Program Director, Advanced General Intelligence for Science Program (AGIS), Transformative Research Innovation Platform of RIKEN platforms (TRIP) Headquarters (Deputy Director, RIKEN Center for Biosystems Dynamics Research (BDR) )
The Hopfield network1 proposed by Dr. Hopfield is closely connected to the complex system of spin glass and is located at the intersection between statistical physics and machine learning. He demonstrated the possibility of modeling the complex world on a computer using artificial neural networks and physical models. Dr. Hinton has also made significant contributions, from the foundations of artificial neural networks to the realization of deep learning. Subsequently, due to improvements in computer performance and the development of deep learning, it has become possible to use machine learning to make predictions about more complex systems, such as the prediction of protein structure using AlphaFold,2 bringing about a revolution in scientific research. In the future, we can expect to see more complex models of the human body and society, and the field of computer-based science will continue to expand. The achievements of these two professors are truly significant.
- Satoshi Matsuoka, Director, RIKEN Center for Computational Science (R-CCS)
As the comments from Program Director Taiji indicate, the achievements that led to the Nobel Prize in Physics being awarded to Dr. John Hopfield and Dr. Geoffrey Hinton were extremely important results that led to a revolution in scientific research. At RIKEN, we are also working on research and development related to AI for Science, and considering the development of the next-generation flagship system “Fugaku Next" (tentative name) to support AI for Science in Japan, but as shown by the Nobel Prize awarded to these two doctors, the movement to develop AI for scientific research and to use AI in scientific research (AI for Science) will rapidly expand around the world.
- Shun-ichi Amari, RIKEN Honorary Science Advisor, Former Director, RIKEN Brain Science Institute
I am very pleased that this year's Nobel Prize in Physics has been awarded to Dr. Geoffrey Hinton and Dr. John Hopfield in the field of artificial intelligence. Physics is a discipline that originally sought to understand the “laws of matter”, but it has now broadened its scope to include the “laws of information”, which could be called the “laws of things”. Indeed, physics has crossed boundaries. The origins of research into artificial intelligence and neural network theory can in part be traced to Japan, and the results of this research have been leveraged internationally to bring us to the age of AI that we live in today.
Starting with the stochastic gradient descent learning method for multilayer neural networks, Dr. Hinton not only made many groundbreaking achievements, including Boltzmann machines and information integration, but also predicted that deep neural networks could perform advanced information discrimination, and achieved groundbreaking results by adding numerous innovations to this. He opened up a new path for artificial intelligence.
Dr. Hopfield proposed the idea of associative memory in neural networks, and led this field by using computer simulations to assess its capacity, attracting many theoretical physicists to the field. The idea of associative memory has been further generalized and is now used as an attention mechanism in large-scale generative models.
Of course, we should not only be impressed by this achievement, but also be concerned about the impact of AI development on society and human civilization.
- 1.Hopfield network
A model of an asynchronous network proposed by Dr. Hopfield, in which there is a symmetrical interaction between units (neurons). It can be applied to neural networks, etc. - 2.AlphaFold
An artificial intelligence program developed by Google's DeepMind that predicts the structure of proteins.